Whoops!

Okay, so here is yet another headline about a medical error, in this case a hospital that managed to kill three premature infants by giving them adult doses of heparin.

Medical errors -- particularly medication errors -- obviously can happen outside of hospitals, but we don't have good data on errors and adverse events in ambulatory care. For hospitals, however, there is a commonly accepted range of estimates, that from 44,000 to 98,000 Americans die every year from avoidable errors made in hospitals. The number who are injured, including many serious injuries (e.g., amputating the wrong leg) is obviously much higher.

You can see where those estimates come from here, which is the first page of the E-book version of the Institute of Medicine's report "To Err Is Human: Building a Safer Health System." The E-book format is kind of dodgy: access to the publication is free, but you can't download it and print it out, you have to look at it on your computer one page at a time. In other words, they're still hoping you'll pay for the printed version. But it's there if you're really interested.

Okay, to kick off this discussion, let me make some basic observations:

Medical intervention, like flying an airplane, is inherently dangerous. You have a long way to fall from the sky, and you also can do a lot of damage by cutting people open, sticking tubes in them, or pumping in or feeding them powerfully bioactive chemicals.

Everybody makes mistakes. If I make a mistake at work (not that it would ever happen), the most dire consequence might be that somebody is sitting around in a conference room wondering where the hell I am, or a questionnaire goes out with an embarassing typo. If a doctor, nurse or pharmacy technician makes a mistake, well .. .

Modern medicine is a very complex undertaking. There are new drugs, new tests, new procedures all the time. New information about risks and counterindications for existing drugs, tests and procedures comes out all the time. It's nearly impossible for anybody to keep track of all the information that might affect patient safety, even in a narrow field. Take my post yesterday about sodium phosphate: most gastroenterologists apparently don't know that it's dangerous for people with kidney failure. That seems pretty basic, but it's also a different specialty.

Historically, the principle method by which medical providers have been made accountable for errors has been malpractice litigation. This does not efficiently discourage errors because:

Mistakes are not the same as malpractice, which requires a finding of negligence. You can make a mistake without being negligent.

Malpractice litigation is an adversarial procedure. It encourages doctors to fight the allegations rather than trying to figure out how to make sure it doesn't happen again. That means trying to suppress information, or interpret it in the most favorable possible light. It means not coming forward in the first place if you know you did something wrong, hoping nobody will notice. It drives physicians to stick together like thieves, creating a culture of cover-up and avoidance.

Malpractice litigation is mostly directed at finding fault in individuals, and getting them and/or their insurance companies to pay up. It doesn't encourage analyzing systems to find ways of making them mistake proof.

So, there is growing interest in systems approaches to medical errors. Try to figure out where the points of vulnerability are that lead to mistakes, and fix the physical environment, the procedures, the job descriptions so that mistakes are impossible to make. For example, people used to be injured by getting hooked up to the wrong kind of gas. Now the fittings for oxygen and anaesthesetics are incompatible. And have you noticed how they always ask your birthday before handing you your prescription? That's to make sure you are the right Pemberton G. Throckmorton.

We'll get into more depth on this later. But in the meantime, remember, even without mistakes, medical intervention is dangerous -- and as a matter of fact, it can be hard to draw the line between mistakes and bad luck. There is an awful lot of judgment involved in trying to trade off risks and benefits. There are deep psychological and philosophical issues in deciding what is appropriate and what is just plain nuts. Hospitals are always going to be very dangerous places.